To retrospectively investigate the feasibility of radiofrequency ablation (RFA) in treating advanced hepatocellular carcinoma (HCC) using standard ultrasound-guided percutaneous RFA. A total of 655 patients with unresectable advanced HCC underwent ultrasound-guided percutaneous RFA therapy at our institution between July 2000 to September 2001. Ninety-two of those patients, representing 136 tumors, were selected for analysis based on the following criteria: presence of UICC/AJCC-TNM (6th edition) stage III and IV advanced HCC, (III: n=82 patients, with 126 tumors; IV: n=10 patients, with 10 tumors); extensive portal vein or inferior vena cava tumor thrombus; extrahepatic metastasis after surgical resection; and complete follow-up data. Follow-up consisted of enhanced computed tomography (CT) performed at one month post-RFA treatment, then every three months. Contrast-enhanced ultrasound (CEUS) was performed in 51 (55.4%) patients before RFA. The standard treatment using optimal strategies were applied in (72.8%) 67 patients. The established strategies included: (1) select RFA indications based on CEUS results; (2) design radical protocols based on invasive range showed by CEUS; (3) multiple overlapping ablations based on mathematical protocols; (4) two or three bipolar RFA electrodes with three-dimensional localization; (5) color ultrasound-guided percutaneous ablation of tumor feeding artery (PAA)/transcatheter arterial chemoembolization (TACE) + RFA for HCC with rich supply. The other 25 patients (27.2 %) were treated with conventional RFA protocols. The ablation procedure was considered a success if no abnormal enhancement or wash-out was detected in the treated area on the CT scan at one month. All patients had received liver protection treatments following RFA. Chi-squared test or Fisher's exact test were used to compare the early complete tumor necrosis rates and the local recurrence rates. Survival was estimated by Kaplan-Meier analysis and log-rank test. P less than 0.05 was considered statistically significant. The RFA-treated tumors ranged in size from 1.5 to 7.0 cm (average: 4.5 cm). Fifty-nine patients had solitary tumor, and the remaining 33 had multiple tumors (2 to 4 tumors). Patients were classified by Child-Pugh score as A (n=58), B (n=32) and C (n=2). Early complete tumor necrosis rate after initial RFA was 90.4% (123/136 tumors). Serious complications developed in two patients (2.2%). No treatment-related death occurred. Follow-up ranged from 3-134 months. Local recurrence rate was 16.9% (23/136 tumors). The 1-, 3- and 5-year overall survival rates were 83.3%, 48.3% and 21.9%, respectively, and the median survival time was 35 months. Stratification analysis indicated the early complete tumor necrosis rate was higher in groups of patients with Child-Pugh A score (98.3%) , CEUS administration (98.0%), and standard treatment (97.0%). The local recurrence rate was lower in groups of patients with tumors less than or equal to 3.0 cm (5.9%), CEUS administration (11.8%), and standard treatment (16.4%). The 5-year survival was significantly higher in patients with Child-Pugh A, tumors less than or equal to 3.0 cm, CEUS administration, and standard treatment (all, P less than 0.05). RFA treatment of patients with advanced HCC, tumors less than 7.0 cm, and without thrombosis in the main vessels was efficacious. The RFA treatment strategy and subsequent liver protection therapy in RFA may improve survival.